University of Washington Regional Epilepsy Center at Harborview Medical Center

Dense Array EEG and Epilepsy

by Mark D. Holmes, M.D.

The evaluation of epileptic seizures requires detailed assessment of the history and clinical examination, and laboratory studies including standard electroencephalography (EEG), brain magnetic resonance imaging (MRI), and, depending upon circumstances, neuropsychological evaluation, and various hematologic and biochemical assays. Despite major advances in neuroimaging, the single most important laboratory tool in the evalution of epilepsy remains the EEG. Nonetheless, standard scalp EEG has some severe limitations. Typically, 16-21 electrodes are applied to the scalp in conventional (international 10-20) recordings. This results in interelectrode distances of several cm, which leads to very poor spatial resolution, so that localization of EEG findings from standard scalp recordings is limited, at best, to gross lobar or hemispheric regions.

Recent technological advances are likely to change this state of affairs and lead to an expanded role for the EEG in epilepsy. One advance is the capability to record from the scalp with a "dense array" of 256 EEG electrodes. With reduced interelectrode distances, spatial resolution is markedly improved, and approaches the theoretical "spatial Nyquist"- the mininum interelectrode distance required to maximize spatial information from scalp recordings. Furthermore, the 256 channel electrode net that is utilized in dense array recordings covers portions of the face and neck, and, in contrast to conventional EEG, permits characterization, as much as is feasible, of electrical activity arising from basal brain regions. Dense array EEG recording is used in conjunction with sophisticated methods of EEG source analysis and realistic models of the head and brain anatomy. Solutions to source analysis are restricted to the cerebral cortex, the brain regions known to generate the EEG, and, by application of a standard MRI model, take into account typical head and brain geometry and the electrical properties of the cranial tissues. Dense array EEG recordings are possible for either short-term 1-2 hour recordings, or when required, for longterm monitoring for up to 48-72 hours.

At the University of Washington Regional Epilepsy Center dense array EEG is being used to study patients with both generalized and localization-related epilepsy syndromes. These studies are leading to novel insights into the neuronal network activated during epileptiform discharges. For example, one study examined patients with typical absence, the prototypic generalized seizure, and suggests that these seizures may not be truly "generalized". Rather, only restricted cortical areas are activated at the onset and during the propagation of the spike-wave bursts. Cortical areas preferentially involved in absence include frontopolar and mesial frontal cortex. Similarly, in a series of patients with juvenile myoclonic epilepsy (JME), a common generalized epilepsy syndrome, highly restricted cortical areas are also found to be active during discharges. In cases of JME, orbitofrontal and temporal cortex are almost always involved, with less common activation of mesial frontal, parietal, and occipital regions. In the future, knowledge of the pathologic neuronal circuitry in refractory generalized seizures may lead to novel treatments.

Dense array EEG may also disclose the complex dynamics of interictal epileptiform discharges in subjects with temporal lobe epilepsy. One important finding is that both temporal lobes frequently contribute sources during the time course of a single interictal spike, even when conventional EEG suggests exclusively unilateral localization. This observation gives credence to the notion that temporal lobe epilepsy is usually a bilateral process. Another finding is that extratemporal regions, such as orbitofrontal cortex, may also be part of the cortical network activated during an apparent temporal discharge. One unanswered question is, does the interictal spike recapitulate the epileptic circuit involved during the clinical seizure? Future research will address this question, and others, in order to place these new results into a useful clinical context.

The primary clinical role for dense array EEG at the Regional Epilepsy Center is as a tool for localizing seizures in patients who are candidates for epilepsy surgery. We have successfully monitored subjects continuously for periods of 24-72 hours and have recorded seizures in nearly all. Our goal is to compare the results of seizure onset and propagation, as predicted by dense array EEG, to standard methods of evaluation, including invasive EEG monitoring.

The outcome in one case of a subject with refractory extratemporal epilepsy forms a basis for optimism that dense array EEG recordings of partial seizures may accurately predict ictal onsets. In this patient, standard EEG recordings disclosed widespread, poorly localized interictal spikes with left posterior quadrant preponderance, while conventional longterm monitoring disclosed seizures that could not be localized. Prior to invasive EEG recordings, dense array EEG studies captured a clinical seizure and source analysis disclosed that the event originated from left posterior inferior occipital cortex. This prediction was confirmed precisely on subsequent invasive recordings. The resection was carried out based on the results of the intracranial studies and the individual has been seizure-free nearly 18 months after the operation. We anticipate that dense array EEG may one day reduce the need for invasive EEG recordings, and at the very least, help guide the placement of intracranial electrodes. In the near future, work with dense array EEG will co-register an individual patient's own MRI (rather than a standard MRI model) to the electrographic data to obtain even more accurate spatial resolution to guide neurosurgical intervention.

The top half of the figure is a topographic dense array EEG map of the initial one second of seizure onset, looking from the top of the head (nose at the top). The bottom half of the figure displays the source analysis of seizure onset, with origin at the left inferior occipital cortex (white voxels indicate maximal intensity).